# Import libraries
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn import model_selection
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
from sklearn.neighbors import KNeighborsClassifier
from catboost import CatBoostClassifier
from sklearn.ensemble import BaggingClassifier
from sklearn.tree import DecisionTreeClassifier
from xgboost import XGBClassifier
from sklearn.metrics import classification_report
import shap
import optuna
from sklearn.metrics import accuracy_score
import matplotlib.pyplot as plt
import warnings
warnings.filterwarnings('ignore')

# Import data
data = pd.read_csv("creditcard.csv")

# Examine the data
data.head()

# Info about the dataset
data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 284807 entries, 0 to 284806
Data columns (total 31 columns):
 #   Column  Non-Null Count   Dtype  
---  ------  --------------   -----  
 0   Time    284807 non-null  float64
 1   V1      284807 non-null  float64
 2   V2      284807 non-null  float64
 3   V3      284807 non-null  float64
 4   V4      284807 non-null  float64
 5   V5      284807 non-null  float64
 6   V6      284807 non-null  float64
 7   V7      284807 non-null  float64
 8   V8      284807 non-null  float64
 9   V9      284807 non-null  float64
 10  V10     284807 non-null  float64
 11  V11     284807 non-null  float64
 12  V12     284807 non-null  float64
 13  V13     284807 non-null  float64
 14  V14     284807 non-null  float64
 15  V15     284807 non-null  float64
 16  V16     284807 non-null  float64
 17  V17     284807 non-null  float64
 18  V18     284807 non-null  float64
 19  V19     284807 non-null  float64
 20  V20     284807 non-null  float64
 21  V21     284807 non-null  float64
 22  V22     284807 non-null  float64
 23  V23     284807 non-null  float64
 24  V24     284807 non-null  float64
 25  V25     284807 non-null  float64
 26  V26     284807 non-null  float64
 27  V27     284807 non-null  float64
 28  V28     284807 non-null  float64
 29  Amount  284807 non-null  float64
 30  Class   284807 non-null  int64  
dtypes: float64(30), int64(1)
memory usage: 67.4 MB

# Check for missing values
data.isna().sum()

Time      0
V1        0
V2        0
V3        0
V4        0
V5        0
V6        0
V7        0
V8        0
V9        0
V10       0
V11       0
V12       0
V13       0
V14       0
V15       0
V16       0
V17       0
V18       0
V19       0
V20       0
V21       0
V22       0
V23       0
V24       0
V25       0
V26       0
V27       0
V28       0
Amount    0
Class     0
dtype: int64

# Descriptive statistics
data.describe().T

# Set the maximum number of rows to display to None (to display all rows)
pd.set_option('display.max_rows', None)

# Grouped statistics by Class
data.groupby(['Class']).describe().T

# Set the Seaborn theme to a whitegrid style
sns.set_theme(style="whitegrid")

# Histograms and boxplots for features
for col in data.columns:
    print(col)
    print('Skew :', round(data[col].skew(), 2))
    plt.figure(figsize = (15, 4))
    plt.subplot(1, 2, 1)
    data[col].hist(grid=False)
    plt.ylabel('count')
    plt.subplot(1, 2, 2)
    sns.boxplot(x=data[col])
    plt.show()

Time
Skew : -0.04

V1
Skew : -3.28

V2
Skew : -4.62

V3
Skew : -2.24

V4
Skew : 0.68

V5
Skew : -2.43

V6
Skew : 1.83

V7
Skew : 2.55

V8
Skew : -8.52

V9
Skew : 0.55

# Distribution of Class
for col in data.columns:
    ax = sns.violinplot(x='Class', y=col, data=data)
    plt.show()

# Undersampling for visualisations
fraud_df = data.loc[data['Class'] == 1]
non_fraud_df = data.loc[data['Class'] == 0].sample(492)

data_vis = pd.concat([fraud_df, non_fraud_df])

# Plot pairplot between numeric columns
sns.set_theme(style="whitegrid")
sns.pairplot(data_vis, height=3, hue="Class", palette='bright')
sns.set(font_scale=0.8)

# Plot correlations
sns.set_theme(style="whitegrid")
corr = data_vis.corr(numeric_only=True)
plt.figure(figsize=(10,10))
sns.heatmap(corr)

<Axes: >

# Plot feature clustering
sns.clustermap(data_vis.corr(numeric_only=True), center=0)

<seaborn.matrix.ClusterGrid at 0x245edffab70>

# Undersampling
fraud_df = data.loc[data['Class'] == 1]
non_fraud_df = data.loc[data['Class'] == 0].sample(20000)

data = pd.concat([fraud_df, non_fraud_df])

# Split data between dependent and independent variables
X = data.drop('Class', axis=1)
y = data['Class']
X_train, _, y_train, _ = train_test_split(X, y, test_size=0.25, random_state=0)

# For testing get full dataset not just subsample
data = pd.read_csv("creditcard.csv")
X_full = data.drop('Class', axis=1)
y_full = data['Class']
_, X_test, _, y_test = train_test_split(X_full, y_full, test_size=0.25, random_state=0)

# Run different models to compare
def run_exps(X_train, y_train, X_test, y_test):
    dfs = []
    results = []
    names = []
    # Define which models to test
    models = [
          ('Logistic Regression', LogisticRegression()),
          ('Random Forest', RandomForestClassifier()),
          ('KNN', KNeighborsClassifier()),
          ('CatBoost', CatBoostClassifier(verbose=False)),
          ('Bagging', BaggingClassifier()),
          ('LDA', LinearDiscriminantAnalysis()),
          ('XGB', XGBClassifier())
        ]

    # Specify scoring metrics and labels
    scoring = ['accuracy', 'precision', 'recall', 'f1', 'roc_auc']
    target_names = ['not-fraud', 'fraud']

    # Iterate through all models
    for name, model in models:
        # Do 3 fold cross validation
        kfold = model_selection.KFold(n_splits=3, shuffle=True, random_state=0)
        cv_results = model_selection.cross_validate(model, X_test, y_test, cv=kfold, scoring=scoring)
        # Fit the model
        clf = model.fit(X_train, y_train)
        # Make predictions
        y_pred = clf.predict(X_test)
        print(name)
        # Print classification repot
        print(classification_report(y_test, y_pred, target_names=target_names))
        # Append to the results
        results.append(cv_results)
        names.append(name)
        this_df = pd.DataFrame(cv_results)
        this_df['model'] = name
        dfs.append(this_df)
        
    final = pd.concat(dfs, ignore_index=True)
    return final
    
final = run_exps(X_train, y_train, X_test, y_test)

Logistic Regression
              precision    recall  f1-score   support

   not-fraud       1.00      1.00      1.00     71082
       fraud       0.34      0.84      0.48       120

    accuracy                           1.00     71202
   macro avg       0.67      0.92      0.74     71202
weighted avg       1.00      1.00      1.00     71202

Random Forest
              precision    recall  f1-score   support

   not-fraud       1.00      1.00      1.00     71082
       fraud       0.79      0.97      0.87       120

    accuracy                           1.00     71202
   macro avg       0.90      0.99      0.94     71202
weighted avg       1.00      1.00      1.00     71202

KNN
              precision    recall  f1-score   support

   not-fraud       1.00      1.00      1.00     71082
       fraud       0.16      0.12      0.14       120

    accuracy                           1.00     71202
   macro avg       0.58      0.56      0.57     71202
weighted avg       1.00      1.00      1.00     71202

CatBoost
              precision    recall  f1-score   support

   not-fraud       1.00      1.00      1.00     71082
       fraud       0.85      0.97      0.90       120

    accuracy                           1.00     71202
   macro avg       0.92      0.98      0.95     71202
weighted avg       1.00      1.00      1.00     71202

Bagging
              precision    recall  f1-score   support

   not-fraud       1.00      1.00      1.00     71082
       fraud       0.67      0.95      0.79       120

    accuracy                           1.00     71202
   macro avg       0.84      0.97      0.89     71202
weighted avg       1.00      1.00      1.00     71202

LDA
              precision    recall  f1-score   support

   not-fraud       1.00      1.00      1.00     71082
       fraud       0.85      0.75      0.80       120

    accuracy                           1.00     71202
   macro avg       0.92      0.87      0.90     71202
weighted avg       1.00      1.00      1.00     71202

XGB
              precision    recall  f1-score   support

   not-fraud       1.00      1.00      1.00     71082
       fraud       0.81      0.97      0.89       120

    accuracy                           1.00     71202
   macro avg       0.91      0.99      0.94     71202
weighted avg       1.00      1.00      1.00     71202

# Pivot the table
results_long = pd.melt(final, id_vars=['model'], var_name='metrics', value_name='values')

# Fit time metrics
time_metrics = ['fit_time', 'score_time'] 

# Get data without fit
results_long_nofit = results_long.loc[~results_long['metrics'].isin(time_metrics)] 
results_long_nofit = results_long_nofit.sort_values(by='values')

# Plot metrics for different models
plt.figure(figsize=(32, 24))
g = sns.boxplot(x="model", y="values", hue="metrics", data=results_long_nofit, palette="Set3")
plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)
plt.title('Comparison of Model by Classification Metric')

Text(0.5, 1.0, 'Comparison of Model by Classification Metric')

# Save metric comparison for tableau visualisation
results_long.to_csv("metric_comparison.csv", index=False)

def objective(trial):
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
    
    # Suggest hyperparameters
    param = {
        "iterations": trial.suggest_int("iterations", 100, 1000),
        "learning_rate": trial.suggest_float("learning_rate", 1e-3, 0.3, log=True),
        "depth": trial.suggest_int("depth", 4, 10),
        "l2_leaf_reg": trial.suggest_float("l2_leaf_reg", 1e-3, 10, log=True),
        "border_count": trial.suggest_int("border_count", 32, 255),
        "loss_function": "Logloss",
        "eval_metric": "AUC",
        "random_seed": 42,
        "verbose": False,
    }

    # Create the classifier model with the suggested hyperparameters
    model = CatBoostClassifier(**param)
    
    # Train the model
    model.fit(X_train, y_train, 
          eval_set=[(X_test, y_test)],
          verbose=False)
    
    # Make predictions
    preds = model.predict(X_test)
    
    # Calculate accuracy
    accuracy = accuracy_score(y_test, preds)
    
    return accuracy

# Create a study object and specify that we want to maximize accuracy
study = optuna.create_study(direction='maximize')
study.optimize(objective, n_trials=20)

# Output the best parameters and the best score
print("Best trial:")
print(study.best_trial.params)

best_params = study.best_trial.params

[I 2025-03-04 19:35:54,704] A new study created in memory with name: no-name-fe208227-ccde-48bb-97a2-6c1bbf5f4b4f
[I 2025-03-04 19:36:01,462] Trial 0 finished with value: 0.995608685045133 and parameters: {'iterations': 885, 'learning_rate': 0.11151189837398591, 'depth': 4, 'l2_leaf_reg': 1.033226659999579, 'border_count': 110}. Best is trial 0 with value: 0.995608685045133.
[I 2025-03-04 19:36:10,576] Trial 1 finished with value: 0.99585264698707 and parameters: {'iterations': 522, 'learning_rate': 0.059267481563127826, 'depth': 8, 'l2_leaf_reg': 0.061449528975144656, 'border_count': 253}. Best is trial 1 with value: 0.99585264698707.
[I 2025-03-04 19:36:19,173] Trial 2 finished with value: 0.9960966089290071 and parameters: {'iterations': 384, 'learning_rate': 0.027880864504313468, 'depth': 9, 'l2_leaf_reg': 0.0065429338048244015, 'border_count': 144}. Best is trial 2 with value: 0.9960966089290071.
[I 2025-03-04 19:36:25,226] Trial 3 finished with value: 0.9939009514515735 and parameters: {'iterations': 749, 'learning_rate': 0.0014388263612415283, 'depth': 4, 'l2_leaf_reg': 0.011563804502100108, 'border_count': 50}. Best is trial 2 with value: 0.9960966089290071.
[I 2025-03-04 19:36:37,605] Trial 4 finished with value: 0.99585264698707 and parameters: {'iterations': 658, 'learning_rate': 0.0011662716924927036, 'depth': 9, 'l2_leaf_reg': 0.2997709039705435, 'border_count': 53}. Best is trial 2 with value: 0.9960966089290071.
[I 2025-03-04 19:36:54,877] Trial 5 finished with value: 0.9965845328128812 and parameters: {'iterations': 686, 'learning_rate': 0.108963912725192, 'depth': 9, 'l2_leaf_reg': 0.13667900463348343, 'border_count': 233}. Best is trial 5 with value: 0.9965845328128812.
[I 2025-03-04 19:36:58,174] Trial 6 finished with value: 0.9946328372773847 and parameters: {'iterations': 352, 'learning_rate': 0.08215863465434856, 'depth': 6, 'l2_leaf_reg': 0.0018793815406910699, 'border_count': 101}. Best is trial 5 with value: 0.9965845328128812.
[I 2025-03-04 19:37:06,862] Trial 7 finished with value: 0.99585264698707 and parameters: {'iterations': 529, 'learning_rate': 0.008776392623469628, 'depth': 8, 'l2_leaf_reg': 0.05298179634995281, 'border_count': 171}. Best is trial 5 with value: 0.9965845328128812.
[I 2025-03-04 19:37:19,143] Trial 8 finished with value: 0.9960966089290071 and parameters: {'iterations': 939, 'learning_rate': 0.013077353100517226, 'depth': 7, 'l2_leaf_reg': 0.0034608713180916194, 'border_count': 241}. Best is trial 5 with value: 0.9965845328128812.
[I 2025-03-04 19:37:21,791] Trial 9 finished with value: 0.9953647231031959 and parameters: {'iterations': 177, 'learning_rate': 0.0019606858361810695, 'depth': 8, 'l2_leaf_reg': 0.5453560426387941, 'border_count': 93}. Best is trial 5 with value: 0.9965845328128812.
[I 2025-03-04 19:38:05,370] Trial 10 finished with value: 0.995608685045133 and parameters: {'iterations': 757, 'learning_rate': 0.28993352338648626, 'depth': 10, 'l2_leaf_reg': 6.850769331795025, 'border_count': 202}. Best is trial 5 with value: 0.9965845328128812.
[I 2025-03-04 19:38:23,379] Trial 11 finished with value: 0.99585264698707 and parameters: {'iterations': 355, 'learning_rate': 0.02970624740403345, 'depth': 10, 'l2_leaf_reg': 0.014374122261444276, 'border_count': 160}. Best is trial 5 with value: 0.9965845328128812.
[I 2025-03-04 19:38:33,064] Trial 12 finished with value: 0.9941449133935106 and parameters: {'iterations': 412, 'learning_rate': 0.28490643037403485, 'depth': 9, 'l2_leaf_reg': 0.013114634404750563, 'border_count': 204}. Best is trial 5 with value: 0.9965845328128812.
[I 2025-03-04 19:38:34,348] Trial 13 finished with value: 0.9968284947548183 and parameters: {'iterations': 111, 'learning_rate': 0.029888067447727866, 'depth': 6, 'l2_leaf_reg': 0.21185146525013773, 'border_count': 205}. Best is trial 13 with value: 0.9968284947548183.
[I 2025-03-04 19:38:36,357] Trial 14 finished with value: 0.9960966089290071 and parameters: {'iterations': 164, 'learning_rate': 0.005363925159225605, 'depth': 6, 'l2_leaf_reg': 0.14502165836997002, 'border_count': 217}. Best is trial 13 with value: 0.9968284947548183.
[I 2025-03-04 19:38:37,497] Trial 15 finished with value: 0.9948767992193218 and parameters: {'iterations': 100, 'learning_rate': 0.0381826994369323, 'depth': 6, 'l2_leaf_reg': 2.459943700511334, 'border_count': 182}. Best is trial 13 with value: 0.9968284947548183.
[I 2025-03-04 19:38:42,784] Trial 16 finished with value: 0.9946328372773847 and parameters: {'iterations': 652, 'learning_rate': 0.14155396703229242, 'depth': 5, 'l2_leaf_reg': 0.12399105606326052, 'border_count': 234}. Best is trial 13 with value: 0.9968284947548183.
[I 2025-03-04 19:38:53,756] Trial 17 finished with value: 0.9948767992193218 and parameters: {'iterations': 839, 'learning_rate': 0.003746260101754989, 'depth': 7, 'l2_leaf_reg': 0.036796550632018976, 'border_count': 221}. Best is trial 13 with value: 0.9968284947548183.
[I 2025-03-04 19:38:57,361] Trial 18 finished with value: 0.9951207611612588 and parameters: {'iterations': 277, 'learning_rate': 0.019335921726515777, 'depth': 7, 'l2_leaf_reg': 1.4738963132173866, 'border_count': 186}. Best is trial 13 with value: 0.9968284947548183.
[I 2025-03-04 19:39:02,828] Trial 19 finished with value: 0.99585264698707 and parameters: {'iterations': 636, 'learning_rate': 0.04802106056081255, 'depth': 5, 'l2_leaf_reg': 0.27403693341976687, 'border_count': 134}. Best is trial 13 with value: 0.9968284947548183.

Best trial:
{'iterations': 111, 'learning_rate': 0.029888067447727866, 'depth': 6, 'l2_leaf_reg': 0.21185146525013773, 'border_count': 205}

# Run CatBoostClassifier
model = CatBoostClassifier(**best_params)
model.fit(X_train, y_train, verbose=0)
model_features = X_train.columns

# Expected columns
for column in model_features:
    if column not in X_train.columns:
        X_test[column] = 0

# Rearange columns
X_test = X_test[model_features]

# Predictions
y_pred = model.predict(X_test)
print(classification_report(y_test, y_pred, target_names=['not-fraud', 'fraud']))

              precision    recall  f1-score   support

   not-fraud       1.00      1.00      1.00     71082
       fraud       0.85      0.87      0.86       120

    accuracy                           1.00     71202
   macro avg       0.92      0.93      0.93     71202
weighted avg       1.00      1.00      1.00     71202

# Create a CatBoostClassifier model
model = CatBoostClassifier(**best_params)
model.fit(X_train, y_train, verbose=0)
y_pred = model.predict(X_test)

# Plot feature importance
explainer = shap.TreeExplainer(model)
shap_values = explainer.shap_values(X_train)

# Plot summary
shap.summary_plot(shap_values, X_train)

	Time	V1	V2	V3	V4	V5	V6	V7	V8	V9	...	V21	V22	V23	V24	V25	V26	V27	V28	Amount
0	0.0	-1.359807	-0.072781	2.536347	1.378155	-0.338321	0.462388	0.239599	0.098698	0.363787	...	-0.018307	0.277838	-0.110474	0.066928	0.128539	-0.189115	0.133558	-0.021053	149.62
1	0.0	1.191857	0.266151	0.166480	0.448154	0.060018	-0.082361	-0.078803	0.085102	-0.255425	...	-0.225775	-0.638672	0.101288	-0.339846	0.167170	0.125895	-0.008983	0.014724	2.69
2	1.0	-1.358354	-1.340163	1.773209	0.379780	-0.503198	1.800499	0.791461	0.247676	-1.514654	...	0.247998	0.771679	0.909412	-0.689281	-0.327642	-0.139097	-0.055353	-0.059752	378.66
3	1.0	-0.966272	-0.185226	1.792993	-0.863291	-0.010309	1.247203	0.237609	0.377436	-1.387024	...	-0.108300	0.005274	-0.190321	-1.175575	0.647376	-0.221929	0.062723	0.061458	123.50
4	2.0	-1.158233	0.877737	1.548718	0.403034	-0.407193	0.095921	0.592941	-0.270533	0.817739	...	-0.009431	0.798278	-0.137458	0.141267	-0.206010	0.502292	0.219422	0.215153	69.99

	count	mean	std	min	25%	50%	75%	max
Time	284807.0	9.481386e+04	47488.145955	0.000000	54201.500000	84692.000000	139320.500000	172792.000000
V1	284807.0	1.168375e-15	1.958696	-56.407510	-0.920373	0.018109	1.315642	2.454930
V2	284807.0	3.416908e-16	1.651309	-72.715728	-0.598550	0.065486	0.803724	22.057729
V3	284807.0	-1.379537e-15	1.516255	-48.325589	-0.890365	0.179846	1.027196	9.382558
V4	284807.0	2.074095e-15	1.415869	-5.683171	-0.848640	-0.019847	0.743341	16.875344
V5	284807.0	9.604066e-16	1.380247	-113.743307	-0.691597	-0.054336	0.611926	34.801666
V6	284807.0	1.487313e-15	1.332271	-26.160506	-0.768296	-0.274187	0.398565	73.301626
V7	284807.0	-5.556467e-16	1.237094	-43.557242	-0.554076	0.040103	0.570436	120.589494
V8	284807.0	1.213481e-16	1.194353	-73.216718	-0.208630	0.022358	0.327346	20.007208
V9	284807.0	-2.406331e-15	1.098632	-13.434066	-0.643098	-0.051429	0.597139	15.594995
V10	284807.0	2.239053e-15	1.088850	-24.588262	-0.535426	-0.092917	0.453923	23.745136
V11	284807.0	1.673327e-15	1.020713	-4.797473	-0.762494	-0.032757	0.739593	12.018913
V12	284807.0	-1.247012e-15	0.999201	-18.683715	-0.405571	0.140033	0.618238	7.848392
V13	284807.0	8.190001e-16	0.995274	-5.791881	-0.648539	-0.013568	0.662505	7.126883
V14	284807.0	1.207294e-15	0.958596	-19.214325	-0.425574	0.050601	0.493150	10.526766
V15	284807.0	4.887456e-15	0.915316	-4.498945	-0.582884	0.048072	0.648821	8.877742
V16	284807.0	1.437716e-15	0.876253	-14.129855	-0.468037	0.066413	0.523296	17.315112
V17	284807.0	-3.772171e-16	0.849337	-25.162799	-0.483748	-0.065676	0.399675	9.253526
V18	284807.0	9.564149e-16	0.838176	-9.498746	-0.498850	-0.003636	0.500807	5.041069
V19	284807.0	1.039917e-15	0.814041	-7.213527	-0.456299	0.003735	0.458949	5.591971
V20	284807.0	6.406204e-16	0.770925	-54.497720	-0.211721	-0.062481	0.133041	39.420904
V21	284807.0	1.654067e-16	0.734524	-34.830382	-0.228395	-0.029450	0.186377	27.202839
V22	284807.0	-3.568593e-16	0.725702	-10.933144	-0.542350	0.006782	0.528554	10.503090
V23	284807.0	2.578648e-16	0.624460	-44.807735	-0.161846	-0.011193	0.147642	22.528412
V24	284807.0	4.473266e-15	0.605647	-2.836627	-0.354586	0.040976	0.439527	4.584549
V25	284807.0	5.340915e-16	0.521278	-10.295397	-0.317145	0.016594	0.350716	7.519589
V26	284807.0	1.683437e-15	0.482227	-2.604551	-0.326984	-0.052139	0.240952	3.517346
V27	284807.0	-3.660091e-16	0.403632	-22.565679	-0.070840	0.001342	0.091045	31.612198
V28	284807.0	-1.227390e-16	0.330083	-15.430084	-0.052960	0.011244	0.078280	33.847808
Amount	284807.0	8.834962e+01	250.120109	0.000000	5.600000	22.000000	77.165000	25691.160000
Class	284807.0	1.727486e-03	0.041527	0.000000	0.000000	0.000000	0.000000	1.000000

Preprocessing¶

Explororatory data analysis (EDA)¶

Modelling¶

	Class	0	1
Time	count	284315.000000	492.000000
	mean	94838.202258	80746.806911
	std	47484.015786	47835.365138
	min	0.000000	406.000000
	25%	54230.000000	41241.500000
	50%	84711.000000	75568.500000
	75%	139333.000000	128483.000000
	max	172792.000000	170348.000000
V1	count	284315.000000	492.000000
	mean	0.008258	-4.771948
	std	1.929814	6.783687
	min	-56.407510	-30.552380
	25%	-0.917544	-6.036063
	50%	0.020023	-2.342497
	75%	1.316218	-0.419200
	max	2.454930	2.132386
V2	count	284315.000000	492.000000
	mean	-0.006271	3.623778
	std	1.636146	4.291216
	min	-72.715728	-8.402154
	25%	-0.599473	1.188226
	50%	0.064070	2.717869
	75%	0.800446	4.971257
	max	18.902453	22.057729
V3	count	284315.000000	492.000000
	mean	0.012171	-7.033281
	std	1.459429	7.110937
	min	-48.325589	-31.103685
	25%	-0.884541	-8.643489
	50%	0.182158	-5.075257
	75%	1.028372	-2.276185
	max	9.382558	2.250210
V4	count	284315.000000	492.000000
	mean	-0.007860	4.542029
	std	1.399333	2.873318
	min	-5.683171	-1.313275
	25%	-0.850077	2.373050
	50%	-0.022405	4.177147
	75%	0.737624	6.348729
	max	16.875344	12.114672
V5	count	284315.000000	492.000000
	mean	0.005453	-3.151225
	std	1.356952	5.372468
	min	-113.743307	-22.105532
	25%	-0.689398	-4.792835
	50%	-0.053457	-1.522962
	75%	0.612181	0.214562
	max	34.801666	11.095089
V6	count	284315.000000	492.000000
	mean	0.002419	-1.397737
	std	1.329913	1.858124
	min	-26.160506	-6.406267
	25%	-0.766847	-2.501511
	50%	-0.273123	-1.424616
	75%	0.399619	-0.413216
	max	73.301626	6.474115
V7	count	284315.000000	492.000000
	mean	0.009637	-5.568731
	std	1.178812	7.206773
	min	-31.764946	-43.557242
	25%	-0.551442	-7.965295
	50%	0.041138	-3.034402
	75%	0.571019	-0.945954
	max	120.589494	5.802537
V8	count	284315.000000	492.000000
	mean	-0.000987	0.570636
	std	1.161283	6.797831
	min	-73.216718	-41.044261
	25%	-0.208633	-0.195336
	50%	0.022041	0.621508
	75%	0.326200	1.764879
	max	18.709255	20.007208
V9	count	284315.000000	492.000000
	mean	0.004467	-2.581123
	std	1.089372	2.500896
	min	-6.290730	-13.434066
	25%	-0.640412	-3.872383
	50%	-0.049964	-2.208768
	75%	0.598230	-0.787850
	max	15.594995	3.353525
V10	count	284315.000000	492.000000
	mean	0.009824	-5.676883
	std	1.044204	4.897341
	min	-14.741096	-24.588262
	25%	-0.532880	-7.756698
	50%	-0.091872	-4.578825
	75%	0.455135	-2.614184
	max	23.745136	4.031435
V11	count	284315.000000	492.000000
	mean	-0.006576	3.800173
	std	1.003112	2.678605
	min	-4.797473	-1.702228
	25%	-0.763447	1.973397
	50%	-0.034923	3.586218
	75%	0.736362	5.307078
	max	10.002190	12.018913
V12	count	284315.000000	492.000000
	mean	0.010832	-6.259393
	std	0.945939	4.654458
	min	-15.144988	-18.683715
	25%	-0.402102	-8.688177
	50%	0.141679	-5.502530
	75%	0.619207	-2.974088
	max	7.848392	1.375941
V13	count	284315.000000	492.000000
	mean	0.000189	-0.109334
	std	0.995067	1.104518
	min	-5.791881	-3.127795
	25%	-0.648067	-0.979117
	50%	-0.013547	-0.065566
	75%	0.662492	0.672964
	max	7.126883	2.815440
V14	count	284315.000000	492.000000
	mean	0.012064	-6.971723
	std	0.897007	4.278940
	min	-18.392091	-19.214325
	25%	-0.422453	-9.692723
	50%	0.051947	-6.729720
	75%	0.494104	-4.282821
	max	10.526766	3.442422
V15	count	284315.000000	492.000000
	mean	0.000161	-0.092929
	std	0.915060	1.049915
	min	-4.391307	-4.498945
	25%	-0.582812	-0.643539
	50%	0.048294	-0.057227
	75%	0.648842	0.609189
	max	8.877742	2.471358
V16	count	284315.000000	492.000000
	mean	0.007164	-4.139946
	std	0.844772	3.865035
	min	-10.115560	-14.129855
	25%	-0.465543	-6.562915
	50%	0.067377	-3.549795
	75%	0.523738	-1.226043
	max	17.315112	3.139656
V17	count	284315.000000	492.000000
	mean	0.011535	-6.665836
	std	0.749457	6.970618
	min	-17.098444	-25.162799
	25%	-0.482644	-11.945057
	50%	-0.064833	-5.302949
	75%	0.399922	-1.341940
	max	9.253526	6.739384
V18	count	284315.000000	492.000000
	mean	0.003887	-2.246308
	std	0.824919	2.899366
	min	-5.366660	-9.498746
	25%	-0.497414	-4.664576
	50%	-0.002787	-1.664346
	75%	0.501103	0.091772
	max	5.041069	3.790316
V19	count	284315.000000	492.000000
	mean	-0.001178	0.680659
	std	0.811733	1.539853
	min	-7.213527	-3.681904
	25%	-0.456366	-0.299423
	50%	0.003117	0.646807
	75%	0.457499	1.649318
	max	5.591971	5.228342
V20	count	284315.000000	492.000000
	mean	-0.000644	0.372319
	std	0.769404	1.346635
	min	-54.497720	-4.128186
	25%	-0.211764	-0.171760
	50%	-0.062646	0.284693
	75%	0.132401	0.822445
	max	39.420904	11.059004
V21	count	284315.000000	492.000000
	mean	-0.001235	0.713588
	std	0.716743	3.869304
	min	-34.830382	-22.797604
	25%	-0.228509	0.041787
	50%	-0.029821	0.592146
	75%	0.185626	1.244611
	max	22.614889	27.202839
V22	count	284315.000000	492.000000
	mean	-0.000024	0.014049
	std	0.723668	1.494602
	min	-10.933144	-8.887017
	25%	-0.542403	-0.533764
	50%	0.006736	0.048434
	75%	0.528407	0.617474
	max	10.503090	8.361985
V23	count	284315.000000	492.000000
	mean	0.000070	-0.040308
	std	0.621541	1.579642
	min	-44.807735	-19.254328
	25%	-0.161702	-0.342175
	50%	-0.011147	-0.073135
	75%	0.147522	0.308378
	max	22.528412	5.466230
V24	count	284315.000000	492.000000
	mean	0.000182	-0.105130
	std	0.605776	0.515577
	min	-2.836627	-2.028024
	25%	-0.354425	-0.436809
	50%	0.041082	-0.060795
	75%	0.439869	0.285328
	max	4.584549	1.091435
V25	count	284315.000000	492.000000
	mean	-0.000072	0.041449
	std	0.520673	0.797205
	min	-10.295397	-4.781606
	25%	-0.317145	-0.314348
	50%	0.016417	0.088371
	75%	0.350594	0.456515
	max	7.519589	2.208209
V26	count	284315.000000	492.000000
	mean	-0.000089	0.051648
	std	0.482241	0.471679
	min	-2.604551	-1.152671
	25%	-0.327074	-0.259416
	50%	-0.052227	0.004321
	75%	0.240671	0.396733
	max	3.517346	2.745261
V27	count	284315.000000	492.000000
	mean	-0.000295	0.170575
	std	0.399847	1.376766
	min	-22.565679	-7.263482
	25%	-0.070852	-0.020025
	50%	0.001230	0.394926
	75%	0.090573	0.826029
	max	31.612198	3.052358
V28	count	284315.000000	492.000000
	mean	-0.000131	0.075667
	std	0.329570	0.547291
	min	-15.430084	-1.869290
	25%	-0.052950	-0.108868
	50%	0.011199	0.146344
	75%	0.077962	0.381152
	max	33.847808	1.779364
Amount	count	284315.000000	492.000000
	mean	88.291022	122.211321
	std	250.105092	256.683288
	min	0.000000	0.000000
	25%	5.650000	1.000000
	50%	22.000000	9.250000
	75%	77.050000	105.890000
	max	25691.160000	2125.870000